Adrenergic Receptor

We identified 12 antihypertensive drug classes in the British National Formulary.²² They were: adrenergic neurone blocking drugs; alpha-adrenoceptor blockers, angiotensin-converting enzyme inhibitors, angiotensin-II receptor blockers, beta-adrenoceptor blockers, calcium channel blockers, centrally acting antihypertensive drugs, loop diuretics; potassium-sparing diuretics and aldosterone antagonists, renin inhibitors, thiazides and related diuretics, and vasodilator antihypertensives. Using the drug substance information, we were able to identify pharmacologically active protein targets and the corresponding genes in the DrugBank database (https://www.drugbank.ca/; version 5.1.1).²³ (link) We then identified SNPs to instrument each protein target using the GTeX project data (Release V7; dbGaP Accession phs000424.v7.p2), which contains expression quantitative trait loci analyses of 48 tissues in 620 donors.²⁴ (link) The full GTEx dataset, which consists of 714 donors, is 65.8% male and 85.2% White. SNPs marked as the ‘best SNP’ for the gene (defined by GTEx as the variant with the smallest nominal p-value for a variant-gene pair) in any tissue were selected for analysis.
To validate the SNPs as instruments for the protein targets of the antihypertensive drugs, we estimated their effect on systolic blood pressure using two-sample Mendelian randomization. The SNP-expression association, extracted from GTEx as described above, was on the scale of a standard deviation change in ribonucleic acid (RNA)-expression levels for each additional effect allele. The SNP–systolic blood pressure association was extracted from the systolic blood pressure GWAS in UK Biobank and represented the standard-deviation change in systolic blood pressure for each additional effect allele. These associations were then used to estimate the effect of the protein target on systolic blood pressure (i.e. the standard deviation change in systolic blood pressure per standard deviation change in RNA-expression levels). SNPs with evidence of an effect on systolic blood pressure were retained for the main analysis. This instrument selection process is presented in Supplementary Figure 1, available as Supplementary data at IJE online.

A total of 18 unique experimental structures and 8 computational models of allosteric modulator-bound class A and B GPCRs were prepared for simulations (Fig. 1A and Supplementary Table 1). The GPCR structures bound by NAMs included the MT7-bound M₁R (PDB: 6WJC)^{19 (link)}, Cmpd-15-bound β₂AR (PDB: 5X7D)^{27 (link)}, AS408-bound β₂AR (PDB: 6OBA)^{29 (link)}, NDT9513727-bound C5AR1 (PDB: 6C1Q)³⁰ , Avacopan-bound C5AR1 (PDB: 6C1R)³⁰ , ORG27569-bound CB₁ receptor (PDB: 6KQI)^{31 (link)}, GTPL9431-bound CCR2 (PDB: 5T1A)^{32 (link)}, NNC0640-bound GLP1R (PDB: 5VEX)^{35 (link)}, PF-06372222-bound GLP1R (PDB: 6LN2)³⁶ , and MK-0893-bound GLR (PDB: 5EE7)^{23 (link)}. Six computational models of NAM-bound GPCRs included the MT7-bound M₂R and M₄R, which were built by aligning the 6WJC PDB structure of M₁R to the 5ZK3^{57 (link)} and 5DSG^{58 (link)} PDB structures of M₂R and M₄R, respectively, and copying atomic coordinates of the atropine antagonist and MT7 NAM, as well as the Cmpd-15-bound α_1B-adrenoceptor (α_1BAR), α_2A-adrenoceptor (α_2AAR), α_2C-adrenoceptor (α_2CAR), and β₁-adrenoceptor (β₁AR), which were built by aligning the 5X7D PDB structure of β₂AR to the 7B6W⁵⁹ , 6KUX⁶⁰ , 6KUW⁶¹ , and 7BVQ^{62 (link)} PDB structures of α_1BAR, α_2AAR, α_2CAR, and β₁AR, respectively, and copying atomic coordinates of the carazolol antagonist and Cmpd-15 NAM. The GPCR structures bound by PAMs included the MIPS521-bound A₁AR (PDB: 7LD3)^{24 (link)}, LY2119620-bound M₂R (PDB: 6OIK)^{17 (link)}, LY2119620-bound M₄R (PDB: 7V68)²⁵ , Cmpd-6FA-bound β₂AR (PDB: 6N48)^{28 (link)}, LY3154207-bound D₁R (PDB: 7LJC)^{33 (link)}, AgoPAM-bound FFAR1 (PDB: 5TZY)²⁰ , INT777-bound GPBAR (PDB: 7CFN)^{34 (link)}, and LSN3160440-bound GLP1R (PDB: 6VCB)^{22 (link)}. Two computational models of PAM-bound GPCRs included LY2119620-bound M₁R, which was built by aligning the 6OIK PDB structure of M₂R to the 6OIJ PDB structure of M₁R^{17 (link)} and copying atomic coordinates of the LY2119620 PAM, and LY32154207-bound D₂ receptor (D₂R), which was built by aligning the 7LJC PDB structure of D₁R to the 7JVR PDB structure of D₂R^{63 (link)} and copying atomic coordinates of the SKF-81297 agonist and LY3154207 PAM.
SWISS-MODEL^{64 (link)} homology modeling was applied to restore missing residues in the GPCR structures and models, particularly in the ECL2, ICL2, and ECL3. Charges of the ligands were listed in Supplementary Table 1. All water and heteroatom molecules except the ligands and receptor-bound ions (including the sodium ion in the 6C1R PDB structure of C5AR1 and zinc ion in the 6LN2 PDB structure of GLP-1 receptor) were removed from the structures. The GPCR complexes were embedded in POPC membrane lipid bilayers and solvated in 0.15M NaCl (Fig. 1B). The AMBER^{65 (link)} force field parameter sets were used for our GaMD simulations, specifically ff19SB^{66 (link)} for proteins, GAFF2^{67 (link)} for ligands, LIPID17 for lipids, and TIP3P⁶⁸ for water, except for the A₁AR simulations as obtained from a previous study^{24 (link)} where the CHARMM36m^{69 (link)} force field parameter set was used.

Rabbit polyclonal anti-ALK7 Ab was generated as described previously (5 (link)). Goat anti–mouse IgG2a, Fc-γ–specific Ab (catalog 33416) and rabbit mAbs against ATGL (catalog 2439S, clone 30A4, RRID:AB_2167953), Smad3 (catalog 9523, clone C67H9, RRID:AB_2193182), phospho-Smad3 (Ser 423/425) (catalog 9520, clone C25A9, RRID:AB_2193207), and UCP-1 (catalog 14670S, clone D9D6X, RRID:AB_2687530) were purchased from Cell Signaling Technology. Mouse monoclonal anti–β-actin Ab (catalog A5316, clone AC-74, RRID:AB_476743) was purchased from MilliporeSigma. Rabbit polyclonal anti–β₃ adrenergic receptor Ab (ab94506, RRID:AB_10863818) was purchased from Abcam. Rabbit monoclonal anti-GAPDH Ab (catalog M171-3, RRID:AB_1059773) and polyclonal anti–HA-tag Ab (catalog 561, RRID:AB_591839) were purchased from MBL. Immunoblotting was performed as described previously (7 (link)).